sticerd.lse.ac.uk/dps/case/cbcb/census1.pdfsticerd.lse.ac.uk/dps/case/cbcb/census1.pdf ·...

CASE-BROOKINGS CENSUS BRIEFS

No.1

The Growth and Decline of Cities and Regions

Ruth Lupton and Anne PowerJuly 2004

INTRODUCTIONThe London School of Economics has a founding commitment tounderstanding the causes of social and economic change. It works to showchanges in patterns of development internationally, whether at a large orsmall scale.Within the UK and in the capital in particular, it tries to keep afinger on the pulse of change and to influence both directly and indirectlythe development of policy. This series of briefs on the 2001 census willpresent findings on population, on changes in the size and distribution ofminority ethnic groups, on tenure and household change and onemployment change, explaining their significance for wider changes. It willalso look at these changes at neighbourhood level, with a particular focus onpoorer neighbourhoods and how they have fared in comparison with theirsurrounding district, city, region, and the country as a whole.This first brieflooks at changes in the distribution of population, focusing on urban andregional growth and decline. It relates these trends to government policy inthe fields of economic growth, distribution of wealth, urban regenerationand social policy.

Changes in population distribution and composition help to shape and areshaped by wider trends both within the country and internationally. Wefocus mainly on cities and built up areas because that is where theoverwhelming majority of the population live, but also because that is thefocus of our work at LSE in the Centre for Analysis of Social Exclusion(CASE). Poverty, deprivation and social policy activity are all heavilyconcentrated within cities and towns. Problems of social exclusion are farmore heavily concentrated in urban than in rural areas and the problemsappear more stubborn and intractable, partly because of their veryconcentration. Urban areas are also central to most economic and culturalactivity, including most higher education. Therefore the strength of urbanareas largely dictates the strength of the overall economy. All of thesereasons make cities and towns of great importance to government,particularly a government committed to eradicating social exclusion, childpoverty and inequality of opportunity.

For a long time British cities have been in decline economically and interms of their populations, and since the early 1970s successivegovernments have focused on attempting to reverse this decline. By 1991,there were signs that population drift was slowing and cities were beginningto recover, but the signs of growth were small - many thought insignificant- and on many counts the decline was continuing.

London School of EconomicsHoughton Street

London WC2A 2AETel: 020 7955 6679

CENTRE FOR ANALYSIS OF SOCIAL EXCLUSIONAn ESRC Research Centre

This is the first in a series of briefs on the 2001 census which willpresent findings on population, on changes in the size and distributionof minority ethnic groups, on tenure and household change and onemployment change. This, and other CASE publications, are availablefrom Jane Dickson at CASE, or can be downloaded from our internetsite: http://sticerd.lse.ac.uk/case

The current government is extremely concerned about theprogress of cities. In 1998, it set up the Urban Task Force,chaired by Lord Richard Rogers, to report on the state ofEngland’s cities and propose measures to reverse declineand promote an “urban renaissance”. The report (DETR1999) was widely hailed as a way forward, recognising thatwithout urban regeneration on a spectacular scale, theBritish economy and social fabric would gradually comeunder threat. Shortage of land, wasteful low densitybuilding, growing polarisation, regional decline, congestionand environmental damage, inner city decline and incipientabandonment of the worst affected neighbourhoods, couldall be tackled through a more coherent approach to cityrecovery.The report was quickly followed by an urban WhitePaper (DETR 2000) and an Urban Summit in 2002, wheregovernment invited every main actor in the urban field totake part in a 3-day affirmation of the growth and recoverypotential of cities.The signals of population and job recoverywere indeed strengthening as many claimed, but the picturewas highly variable. It was impossible to judge how muchfurther the outward drift of population and the stronglypolarizing trends of the 1980s had slowed or intensified.TheCensus 2001 findings shed new light on cities and on thechanges that were being celebrated at the Urban Summit.They allow us to quantify the urban and regional changesthat we are experiencing and to understand the importanceof economic, household and housing patterns alongsidechanges in population and ethnic composition.

The government is committed to follow up with a secondUrban Summit in February 2005 and to publish a reportin time for it on the state of our cities, following apreliminary report in 2000.This document on populationchange in Great Britain, based on the UK Census 2001,is the first in a series of Census Briefs produced by CASEthat aim to help advance the debate on the future of citiesand towns.They are inspired by the work of the BrookingsInstitution in the United States whose Center on Urbanand Metropolitan Policy has played a creative role ininforming and in part helping shape the recovery of UScities. By understanding and explaining the causes anddirections of change, key actors can in fact also improvepolicy and influence outcomes.

In the United Kingdom, cities have neither declined as faras in the US nor has government ignored for so long thesigns of decline. Nonetheless, in this country, we havelittle to be complacent about. Many studies show thechronic state of cities, the maldistribution of wealth andjobs, the growing ethnic concentrations in some of themost declining urban neighbourhoods, the regionalimbalances and threat of social unravelling that somecities face (Lupton 2003, Mumford and Power 2003).This brief discusses population changes between 1991and 2001 in the context of intense policy interest, bothhere and in Europe and the United States.

We owe a debt of thanks to Bruce Katz and othercolleagues in Brookings for inspiring us to undertakethis series jointly with them; to Professor William JuliusWilson of Harvard for his constant interest in our workon poor neighbourhoods and his willingness to join thewider urban debate in this country as well as in the US;to Professors Tony Champion, Duncan McLennan andIvan Turok for their challenging advise and willingnessto share expertise; to David Lunts, head of the Urban

Unit at ODPM and the many other colleagues ingovernment who have encouraged us to do this work;also to Richard Best at the Joseph RowntreeFoundation for supporting our original work on theslow death of great cities and to Richard Rogers forlending his expertise and experience to our follow-upto the Urban Task Force, Cities for a Small Country(Rogers and Power 2000).Throughout, we draw on ourwork in CASE for the area study funded by the ESRCwhere we track 12 of the poorest urban areas in thecountry over 7 years, written up by Ruth Lupton(2003) and Katharine Mumford and Anne Power(2003), and on the work of our colleagues at the Centeron Urban and Metropolitan Policy at Brookings, whosework on the US Census can be found ath t t p : / / w w w. b r o o k i n g s . e d u / e s / u r b a n / i s s u e s /demographics/demographics.htm.

ANALYSING POPULATION CHANGE USING THE CENSUSThe 2001 Census was conducted in England and Walesby The Office for National Statistics (ONS) and inScotland by the General Register Office for Scotland(GRO). Data is Crown Copyright and is reproducedhere with the permission of the Controller of HMSOand the Queen’s Printer for Scotland. It provides thefirst major opportunity for ten years to examine changesin Great Britain’s socio-economic geography, down tosmall area level.

However, as is widely known, comparing the 2001 Censuswith results from 1991 is not straightforward. The twoCensuses differ in two important ways. First, the 1991Census results included fewer adjustments for under-enumeration. They did include individuals andhouseholds imputed to exist by enumerators but who hadnot filled in forms, but they were not adjusted to takeaccount of people who had been missed altogether byenumerators. Estimates of the numbers of such ‘missingpeople’ were subsequently derived from a post-Censussurvey and from rolling forward administrative data fromthe 1980s, and were used to feed into later populationestimates. However, the original Census figures were notaltered. The 2001 Census figures, by contrast, wereadjusted before publication to incorporate estimates ofunder-enumeration. They are therefore higher than the1991 Census figures, particularly in urban areas whereunder-enumeration was thought to be highest.

Second, there was an important change to the waystudents were enumerated. Data in 1991 was collectedon a ‘usual address’ basis, with students counted at theirvacation address rather than their term address. In 2001,they were counted at their term address. In majoreducational centres, this had the effect of making 2001Census figures seem high by comparison with 1991 figures.

Both of these difficulties can be got around at the localauthority level by comparing mid-year estimates ofpopulation (MYEs) rather than Census figures. MYEs arecalculated in each Census year, based on the Census andany post-Census adjustments, and rolled forward each

July 2004 • CASE-Brookings Census Briefs • No.12

year between Censuses on the basis of estimates of naturalgrowth and migration. The MYE for 1991 took intoaccount estimates of under-enumeration as well as post-Census adjustments which counted students at their termaddresses rather than their vacation addresses. In thisbrief, which looks at overall population trends for localauthority areas, we therefore compare MYEs, rather thanraw Census data.1

Further complications have subsequently arisen. Censusresults for 2001 showed the overall population to beconsiderably smaller than had been anticipated fromrolled-forward MYEs based on the 1991 Census.Investigation by ONS revealed that this was due to over-estimation in the process of rolling forward the MYEsduring the 1990s. It was thought that ONS hadovercompensated for under-enumeration in the 1991Census when calculating the 1991 MYE, and had alsounderestimated the level of international emigration duringthe decade. Most of the discrepancy affected the counts foryoung males, and affected large urban centres more thansmaller settlements. As a result, ONS and GRO revisedtheir population estimates for the years 1991-2001.However, even after this adjustment had been made,another problem came to light. It became apparent thatpart of the original discrepancy between rolled-forwardMYEs and 2001 Census had come about not just becausethe MYEs were too high but because the 2001 Censuscounts were too low. In other words, there had been ahigher level of under-enumeration in the 2001 Census thanoriginally thought. As a result, further revised 2001estimates were issued in September and November 2003.For many local authorities, these changes were negligible.However, for some large urban authorities, they weresignificant. Manchester was the most extreme case. Itsestimated population went up about 27,000 from the post-Census 2001 estimate to the revised (Nov 2003) estimate.

There may well be further revisions to the figures, sinceONS is still investigating the source of the problems.However, at the time of writing, the November 2003revisions are regarded as the best estimates that can beproduced. We use them here, in comparison with revised1991 estimates, to describe population change.

Further details about the comparability of 1991 and 2001data and the problems that have been encountered with2001 population estimates can be found on the NationalStatistics website (www.statistics.gov.uk). One furthermethodological point to make here concerns theapproach we have taken to identify and compare types ofareas. Primarily to enable comparison with changesduring the 1980s, we have made use of a local-authoritybased classification2 of areas developed by TonyChampion and colleagues at the University of Newcastleand originally used to report on urban trends in Englandusing 1981 and 1991 Census data (Atkins et al. 1996).The classification divides local authorities into types suchas ‘districts with industrial areas’, ‘resort, port andretirement areas’ and ‘principal metropolitan cities’. Sincethe original work in the mid 1990s, it has been revised toincorporate boundary changes during the 1990s and toinclude Scotland and Wales. It is the revised classificationthat is used here and we are very grateful to TonyChampion for making it available to us.

THE DISTRIBUTION OF POPULATION IN GREAT BRITAINSince this series has been developed partly for aninternational readership, we begin by making somepreliminary observations about the population geographyof Great Britain.

England, Scotland and Wales have between them apopulation of about 57.3 million. They have one verylarge city, London, with approximately 7.3 million peopleor about 13% of the entire population. Outside London,there are seven large conurbations: Tyne and Wear,Merseyside, Clydeside, Greater Manchester, WestMidlands, West Yorkshire and South Yorkshire. These arethe settlements built up around the major cities of theindustrial revolution, in Scotland and the North andMidlands of England. Clydeside contains the city ofGlasgow and its hinterland, including districts like NorthLanarkshire and Renfrewshire; Tyne and Wear containsthe city of Newcastle and surrounding districts likeGateshead and North Tyneside; West Yorkshire the citiesof Leeds and Bradford and districts like Calderdale;WestMidlands the city of Birmingham and districts likeCoventry, Sandwell and Dudley; Merseyside the city ofLiverpool and districts like Sefton and Knowlsey; GreaterManchester the principal city of Manchester, the city ofSalford and surrounding districts like Bury, Bolton,Wigan and Trafford; and South Yorkshire the city ofSheffield and surrounding districts like Rotherham.Theseconurbations have populations between about 1 millionand 2.5 million, making a total of about 12.5 million.Themajor cities within them are described in the classificationwe use as the ‘metropolitan cities’, and the other localauthority districts within them as ‘other metropolitandistricts’. In this paper, we discuss changes in both theconurbations as a whole and the principal metropolitancities and other metropolitan districts within them.

Cities outside the major conurbations are described in theclassification we use as non-metropolitan. There are 14large non-metropolitan cities, places such as Edinburgh,Cardiff, Swansea, Bristol and Nottingham, withpopulations of between about 150,000 and 450,000.These are typically large industrial centres or ports. Wealso have 17 ‘small non-metropolitan cities’, withpopulations between 80,000 and 180,000, which areeither smaller industrial centres (such as Middlesbrough),educational centres (like Oxford and Durham), and/ orthe major urban settlements serving largely ruralhinterlands (such as Worcester).3

Map 1 shows the distribution of these cities, and Table 1their populations. Maps 2-8, later in this paper, show thelocal authorities within the conurbations, and theirpopulation sizes, and may be a useful reference at thispoint for those unfamiliar with the geography.

1 Further papers in the series, looking at specific topics such as housing andemployment, are based on Census counts.

2 We acknowledge that there are problems with using a classification based on LAboundaries. Some cities have loosely drawn administrative boundaries, incorporatingouter suburban and rural areas, while others are tightly drawn, making comparisonsdifficult. Census data aggregated to urban areas will make a more sensitiveclassification possible for these kinds of areas, but has not yet been published, and alsogives incomplete coverage, because non-urban areas are not included. We therefore usea local authority-based classification for the time being, while recognising itslimitations.

3 Brighton and Hove, which was created as new local authority in the mid-1990s by theamalgamation of Brighton with Hove, has retained Brighton’s classification of ‘smallnon-metropolitan city’ even though its population considerably exceeds that of theothers in its category.

THE GROWTH AND DECLINE OF CITIES AND REGIONS 3

CITIES IN BRITAIN


MAP 1:

Altogether, about 26 million people (45% of thepopulation) lives in London, the seven otherconurbations and 31 other cities. The remaining localauthority districts, which are not classified as cities,include smaller industrial towns outside the metropolitanareas, seaside resorts and smaller ports, towns inpredominantly rural areas, and remoter rural areas.Table2 shows the full range of district types and thedistribution of population between them.4

It is the cities that are the focus of our work. We firstconsider trends in the 1990s, then look at patterns ofcontinuity and change over the 1980s and 1990s.We thenexamine data at the regional level, raising the question ofwhether changes between district types merely reflectregional level changes, before drawing conclusions abouttrends in the growth and decline of cities and regions.

THE 1990S: LONDON GROWING RAPIDLY,OTHER LARGE CITIES IN DECLINE The population of Great Britain grew by about one and ahalf million people, or 2.7% of its 1991 population,between 1991 and 2001.

The population of cities, taken together, also grew, by244,000, a 1% growth rate (i.e. less than half thegrowth rate of Great Britain as a whole). However, thisfigure masks important differences between cities.London grew very rapidly, especially Inner Londonwhich was the fastest growing district type in thecountry. The capital gained just under half a millionpeople (479,000) in the 1990s, a 7% gain. Meanwhile

Source: ONS : 2001 MYEs (Nov 2003)

* Major conurbations are the principal metropolitan cities and theirsurrounding metropolitan districts

TABLE 1: Populations of Cities in Great Britain

2001 population

(thousands)

London 7308

Major Conurbations* 12542

West Midlands 2570

Greater Manchester 2513

West Yorkshire 2084

Clydeside 1666

Merseyside 1366

South Yorkshire 1266

Tyne and Wear 1078

Large Non-Metropolitan Cities 3629

Edinburgh 449

Bristol 384

Cardiff 307

Leicester 283

Nottingham 269

Kingston upon Hull 243

Plymouth 241

Stoke-on-Trent 240

Swansea 224

Derby 223

Aberdeen 212

Southampton 220

Portsmouth 188

Dundee 145

Small Non-Metropolitan Cities 2202

Brighton and Hove 250

York 181

Bath and North East Somerset 169

Reading 145

Newport 138

Oxford 136

Middlesbrough 136

Preston 131

Norwich 123

Exeter 111

Cambridge 110

Cheltenham 110

Gloucester 110

Worcester 93

Durham 88

Stirling 86

Lincoln 86


TABLE 2: Local Authority District Types, and their Populations

Population % ofNumber of 2001 Overall % of

District Type districts (thousands) population districts Example

London 33 7308 13% 8% Lewisham,Brent

Principal met city 7 3915 7% 2% Manchester,Glasgow

Other met 36 8627 15% 9% Knowsley,borough RotherhamLarge non-met 14 3629 6% 3% Bristol,city EdinburghSmall non-met 17 2202 4% 4% Oxford,city DurhamDistrict with 71 7866 14% 17% Slough,industrial areas BlackburnDistrict with 24 2950 5% 6% Redditch,new towns Milton KeynesResort, port 31 3447 6% 8% Blackpool,and retirement BournemouthUrban and mixed 98 10798 19% 24% Chester,urban-rural SevenoaksRemoter 77 6621 12% 19% Cotswold,mainly rural Perth and

KinrossGreat Britain 408 57363

Source: ONS : 2001 MYEs (Nov 2003)

4 It will be obvious to a US readership that the terms used to classify cities in this paperare not the same as those commonly used in US commentaries. Perhaps the closestcomparisons can be drawn between conurbations in the UK and ‘metropolitan areas’or ‘metros’ in the US. For a detailed consideration of the similarities and differencesbetween US and UK definitions, see Tunstall (2004).

the large industrial conurbations of the North,Midlands and Scotland lost, between them, about270,000 people, 2.1% of their population at the start ofthe decade. Outside London, all of the largeconurbations except West Yorkshire lost population.Thus while London and the other conurbations,together, showed net growth of over 200,000 people,this was almost entirely due to the large growth ofLondon outstripping losses elsewhere (Table 3).

The fortunes of the individual conurbations outsideLondon also varied from one another. West Yorkshireactually gained population (although a lot more slowlythan the national average) whilst among the losers,Merseyside, Tyne and Wear and Clydeside did muchworse than the West Midlands, Greater Manchester andSouth Yorkshire.

Other large cities also lost population, although on asmaller scale, while small cities gained population atabout the national rate. Industrial areas lost population,while rural areas, mixed urban and rural areas, coastaltowns, and districts with new towns all grew faster thanthe national average (Table 4).

CLASSIFYING RATES OF POPULATION CHANGEUsing an adapted version of a classification applied to UScities by Glaeser and Shapiro (2003), we have groupeddistricts according to their rate of population change inthe 1990s. Glaeser and Shapiro use a four-wayclassification:

• High flyers: Cities growing by 10% or more

• Modest Growers: Cities growing by between 2% and 10%

• Unchanged: Cities changing by between -2% and +2%

• Decliners: Cities declining by 2% or more

Because of the less dramatic population changes inEnglish districts, we have adapted this classification andre-oriented it around the overall population growth rate(about 3%) in Great Britain to refer to districts growingby 3%-10% as ‘growers’ rather than ‘modest growers’ andredefined the ‘unchanged’ category as ‘in the balance’,further subdividing it to distinguish between slightdecliners, those with minimal change and slight growers.This gives us a six-way classification, as follows:

• High flyers: Districts growing by 10% or more

• Growers: Districts growing by between 3% and 10%

• In the Balance: Slight growers: Districts growing bybetween 0.5% and 3%

• In the Balance: Minimal change: Districts changingbetween -0.5% and 0.5%

• In the Balance: Slight Decliners: Districts declining bybetween -0.5% and -3%

• Decliners: Districts declining by 3% or more

Table 5 shows that 44% of all districts were ‘growers’with a further 12% high flyers. Thus 56% had growthrates exceeding the approximate national rate of increase.Only 7% were in decline (more than 3% loss) and of the37% of districts that were ‘in the balance’, 16% showedgrowth, while 13% showed decline.The overall picture isone of slightly greater growth than decline. However,with the exception of London, the general pattern wasthat less urban the district, the more likely it was to grow.Over two-thirds of resort, port and retirement districtsand mixed urban and rural districts grew by more thanthe national rate, along with more than four-fifths ofrural districts.


TABLE 3: Population Changes in the Conurbations 1991-2001

%1991 2001 Change change

population population 1991- 1991-(000s) (000s) 2001 2001

London 6829 7308 479 7.0Inner London 2599 2838 239 9.2Outer London 4230 4470 240 5.7

Other conurbations 12812 12542 -270 -2.1West Midlands 2619 2570 -49 -1.9Greater Manchester 2553 2512 -41 -1.6West Yorkshire 2062 2084 22 1.1Clydeside 1728 1666 -62 -3.6Merseyside 1438 1366 -72 -5.0South Yorkshire 1288 1266 -22 -1.7Tyne and Wear 1124 1078 -46 -4.1

All conurbations 19642 19850 209 1.1

Source: ONS : 1991 (revised) and 2001 MYEs (Nov 2003)

TABLE 4: Population Changes for District Types outside theConurbations 1991-2001

1991 2001 Change % changepopulation population 1991- 1991-

(000s) (000s) 2001 2001

CITIES (excluding the main conurbations)Large non-met city 3664 3629 -35 -1.0Small non-met city 2131 2202 71 3.3OTHER DISTRICT TYPESDistrict with industrial areas 7732 7866 134 1.7District with new towns 2829 2950 121 4.3Resort, port and retirement 3320 3447 127 3.8Urban and mixed urban-rural 10277 10798 521 5.1Remoter mainly rural 6237 6621 383 6.1Great Britain 55831 57363 1532 2.7

Source: ONS : 1991 (revised) and 2001 MYEs (Nov 2003)

CHANGES IN THE CONURBATIONS AND OTHER CITIESLooking more closely at trends in the conurbations andother cities (Figure 1), we can see big differences betweenthe city types.

FIGURE 1: Population Change Rates in the Cities

The majority of districts in London had growth in excessof the national average, and nearly one third of them camein the high flyer category, principally districts in innerLondon (such as Newham,Tower Hamlets, Hackney andSouthwark).The population of Tower Hamlets grew by anastonishing 21.2% in the decade, and Newham’s by15.6%.5 The City of London saw growth of 37%, from avery low base.

About half of the small non-metropolitan cities also grewsignificantly - places like Worcester, Reading, Gloucester,Exeter and Oxford, and only one in this category(Middlesbrough), was a decliner.

This situation can be contrasted with the position in themajor conurbations outside London and in the otherlarge cities. In the large conurbations (the principalmetropolitan cities and their surrounding districts), therewere no ‘high flyers’ and only one ‘grower’. Of themetropolitan cities themselves, Manchester (-3.3%),Newcastle (-5.1%) Liverpool (-7.0%) and Glasgow (-8.0%) all declined, while Leeds, Sheffield andBirmingham were all ‘in the balance’, Leeds showingslight growth (1.3%) and Sheffield (-1.3%) andBirmingham (-1.9%) slight decline. The surroundingmetropolitan districts followed broadly the same pattern.Districts in Greater Manchester, Merseyside, Clydesideand Tyne and Wear generally did less well than districts inWest Yorkshire, and to a lesser extent in South Yorkshireand the West Midlands. But within the decliningconurbations, the cities themselves did worse than thesurrounding districts. These patterns are illustrated inMaps 2-8.


TABLE 5: Classification of Population Change by Type of District

Inner LondonOuter LondonPrincipal met cityOther met boroughLarge non-met citySmall non-met cityDistrict with industrial areasDistrict with new townsResort, port and retirementUrban and mixed urban-ruralRemoter mainly ruralTotal% of all districts

Decliners (more than

-3% decline)

485163111

307%

Slight decliners (-0.5% to -3%)

12

1631

182135

5213%

Minimal change (-0.5%-0.5%)

1

711632

102

338%

Slight growers(0.5%-3%)

211434

1546

171067

16%

Growers (3%-10%)

414

129

228

195841

17844%

High flyers (more

than 10%)

82

14429

1848

12%

Total

1419

73614174124319877

408

Number of Districts in Each Population Change Category

In the balance

Type of District

5 Although the estimates we use here are the best available, it is worth bearing in mind thatthere are particular problems with obtaining accurate Census counts in local authorityareas such as these with high ethnic minority concentrations, younger than averagepopulations, high proportions of recent immigrants and asylum seekers and highproportions of rented accommodation. It is not inconceivable that some of the changesreported are accounted for by underestimation in 1991, nor indeed that increases havebeen under-estimated because of under counting in 2001. In short, the nature of InnerLondon means that all population estimates have to be treated a little more cautiously thanthey do in areas of less mobile population.


MAP 4

MAP 5

MAP 6

MAP 8

MAP 2

MAP 3

MAP 7

Greater Manchester ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

Merseyside ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

Tyne and Wear ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

West Midlands ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

South Yorkshire ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

West Yorkshire ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

Clydeside ConurbationPopulation of Districts 1991 and 2001 (000s) and % change

Population change 1991-2001Decliners (more than 3% decline)Slight Decliners (-0.5% to -3%)Minimal Change (-0.5% to 0.5%)Slight Growers (0.5% to 3%)Growers (3% to 10%)High Flyers (10%+)

Population (000’s)19912001


CONTINUING TRENDSThese trends were not good news for the large cities.However they did not represent a new development.Broadly, speaking the trends of the 1990s were acontinuation of the trends of the 1980s (see Table 6).District types that were doing well in the 1980scontinued to do well in the 1990s. Large cities outsideLondon continued to decline, although slightlyslower. Small cities continued to grow, and in factgrew more quickly than in the 1980s. Rural andmixed/urban rural areas continued to grow at muchthe same rate as in the 1980s (more than double thenational average) while growth rates slowed in the newtowns and resort, port and retirement areas, butremained higher than average.

The only dramatic change of direction was in London. Inthe 1980s, London had just about held its population,with a decrease in outer London being offset by anincrease in inner London of about the national average.In the 1990s, as we have seen, the large population

Among the large non-metropolitan cities, there were no ‘high flyers’ and only two ‘growers’, Southampton and Cardiff,although Edinburgh at 2.9% was closely behind. Figure 2 shows the rates of change for the non-metropolitan citiesindividually, showing the different fortunes of the large cities compared with the small cities.

TABLE 6: Population change in the 1990s compared with the 1980s

Inner LondonOuter LondonPrincipal met cityOther met boroughLarge non-met citySmall non-met cityDistrict with industrial areasDistrict with new townsResort, port and retirementUrban and mixed urban-ruralRemoter mainly ruralGreat Britain

198125504255426289363705209476322633310298085837

54815

1991259942304044876936642131773228293320

102776237

55831

2001283844703915862736292202786629503447

107986621

57363

Change1981-1991

49-25

-219-167-413799

195218469401

1017

Change1991-2001

239240

-128-142-3571

134121127521383

1532

% change1981-2001

1.9%-0.6%-5.1%-1.9%-1.1%1.8%1.3%7.4%7.0%4.8%6.9%1.9%

% change1991-2001

9.2%5.7%

-3.2%-1.6%-1.0%3.3%1.7%4.3%3.8%5.1%6.1%2.7%

Population (thousands)

Source: ONS : 1981, 1991 and 2001 MYEs (Nov 2003)

FIGURE 2: % Population change 1991-2001 for Non-Metropolitan Cities

increases throughout the conurbation, and particularly inInner London, put London on a completely differenttrajectory to the other cities (see Figure 3).

FIGURE 3: Population Turnaround in London Compared with other cities


Source: ONS : 1991 and 2001 MYEs (Nov 2003)

Table 7 shows the conurbations and cities individually,pointing to some variations within the overall patterns.The conurbations, with the exception of London, wereall in decline in the 1980s. West Yorkshire was doing theleast badly, and actually showed a slight turnaround inthe 1990s while the other conurbations continued todecline. Most slowed their decline, but Tyne and Wearactually lost proportionally more population in the1990s than the 1980s.

Similarly, there were some gainers among the large citiesin the 1990s, with Portsmouth, Edinburgh and to a muchgreater extent Southampton turning 1980s losses into1990s gains. However, there were also gainers in the1980s that failed to maintain their gain in the 1990s(Nottingham, Aberdeen and Swansea) and losers in the1980s like Hull and Stoke-on-Trent that went into worsedecline in the 1990s.

The small cities showed a more consistent pattern ofgrowth. A majority were on the up in the 1980s andcontinued that trend in the 1990s, being joined by anumber of others like Brighton, Bath, Stirling, Oxford

and Reading that all started to do better in the 1990s thanthey had done previously. However, Norwich, and to amuch greater extent, Middlesbrough, were in steadydecline throughout the period.

City size, then, does not entirely explain patterns ofpopulation change.While large cities have generally beingdoing badly for two decades, and small cities increasinglywell, some cities are bucking the trends in both directions.

DECLINING CITIES OR DECLINING REGIONS?The analysis thus far in this paper, based on localauthority type, shows strong differences between types.However, this does not necessarily mean that district typeis the cause of different trajectories - for example thatlarge cities are doing badly because they are large citiesand are suffering from counter-urbanisation trends. Itmay be the case that district-level trends are driven bywider regional characteristics, such as industrial structureor peripheral location, which merely manifest themselvesin district-type analyses because of the types of districtsthat exist in each region. To explore this possibility, wealso examine trends at the regional level.

Great Britain has eleven administrative ‘regions’, of whichone is the whole of Scotland and one the whole of Wales.6

These are listed in Table 8.These have different physical,economic and social characteristics. At the crudest level,regions in the North and Midlands are largely urban andindustrialised, while regions in the South have a morerural and small town tradition. These differences arereflected in their administrative structures at the localauthority district level. In relation to the classification wehave used throughout this paper, the regions in the Southand East of the country contain mainly small rural andmixed urban/rural districts, while the Northern regionscontain all of the principal metropolitan cities and theirsurrounding districts and a high proportion of industrialareas. (Figure 4).7


TABLE 7: Population Trajectories for Major Urban Districts1980s and 1990s

% change % change 1981-1991 1991-2001

CONURBATIONSLondon 0.4% 7.0%Tyne and Wear -2.8% -4.3%Merseyside -5.5% -5.0%Greater Manchester -2.5% -1.6%West Midlands -2.0% -1.9%West Yorkshire -0.2% 1.1%South Yorkshire -2.2% -1.7%

LARGE NON-MET CITIESCardiff 3.5% 3.5%Derby UA 2.6% 0.1%Nottingham UA 0.4% -3.7%Swansea 0.2% -2.7%Leicester UA -0.6% 0.6%Aberdeen 0.8% -1.0%Plymouth UA -0.8% -4.1%Stoke-on-Trent UA -1.2% -3.6%Edinburgh -2.2% 2.9%Bristol, City of UA -2.2% -2.2%Portsmouth UA -2.4% 0.7%Southampton UA -2.4% 7.3%Kingston upon Hull, City of UA -3.8% -7.6%Dundee -8.3% -6.5%

SMALL CITIESLincoln 9.8% 1.9%Worcester 8.0% 12.1%Cambridge 5.7% 3.2%York UA 4.2% 5.2%Exeter 4.2% 6.1%Cheltenham 4.1% 2.8%Gloucester 3.2% 6.2%Preston 2.8% 0.4%Newport 2.3% 1.6%Brighton and Hove UA 1.4% 4.0%Bath and North East Somerset UA 1.0% 3.7%Stirling 0.7% 6.6%Norwich -0.9% -2.0%Oxford -1.1% 5.3%Reading UA -1.9% 7.5%Durham -2.1% 2.1%Middlesbrough UA -3.9% -6.4%


6 These are the regions at Level 1 of the Nomenclature of Units for Territorial Statistics(NUTS) hierarchy, created by the European Office for Statistics to enable a standardapproach to spatial units across the EU. Level 1 is the first level below the nationallevel. In England, NUTS1 areas correspond to Government Office Regions. The wholeof Scotland is a NUTS1 area, as is the whole of Wales.

7 Figure 4 also highlights various anomalies in the administrative geography of regionsthat affect district type analyses. For example, London is the only principal city that isdivided into separate local authorities. All the others have just one large authority.Yorkshire and Humberside, which is highly industrialized, has only one ‘district withindustrial areas’, because its industrial areas are contained within its two largeconurbations, West and South Yorkshire, and thus classified as metropolitan districts,whereas the East Midlands has no large conurbation and a high proportion of smallerdistricts classified as industrial.

FIGURE 4: Types of District by Region


Analysis of population change at the regional level showsthat in both the 1980s and 1990s, regions in the southand east of the country did better in terms of populationthan regions in the north and west (Figure 5). In the1990s, which is our main focus, London’s populationgrew by 7%, the South East, East and South West byabout 5% each and the East Midlands by 4% (all higherthan the national figure). Meanwhile the North East andNorth West and Scotland actually lost population and theWest Midlands, Yorkshire and Humberside and Waleswere relative losers, gaining population more slowly thanthe national rate.

With the exception of London (which had a rapidturnaround in the 1990s as we have already observed),regions that were doing well in the 1990s generallycontinued to do relatively well, while those that had beendoing badly in the 1980s continued to do badly. Withinthis broad pattern, the South West, North East and Walesdid slightly less well in the 1990s than 1980s, andScotland did slightly better (although still declining).The different sizes of the regions combined with theirgrowth rates to influence what share of overallpopulation change they accounted for. With their largepopulations and high growth rates, London and theSouth-East accounted for over half the total growth inthe 1990s (Table 8, final column). 75% of the growthwas accounted for by what is normally regarded as theprosperous South and East of the country - London, theSouth-East and the East of England.

The question is whether these changes are merely amanifestation of the different composition of the regionsin terms of types of districts within them, and thusexplained by shifts of population between district types,or whether they reflect regional drivers of change.

There is some evidence to support both arguments. On theone hand, it is clear that growing district types wereprimarily in growing regions. Figure 6 shows a verysimplified representation of the situation. The regions aredivided into those with substantial growth (South East,South West, London, East of England and East Midlands),those with some growth (West Midlands, Yorkshire andHumber and Wales), and declining regions (North East,North West, Scotland). District types are similarly dividedaccording to their overall population change - substantial

growers (London, small cities, urban/rural, resort/port,rural and new towns), those with some growth (industrialareas) and decliners (met districts and large non metcities). The graph shows that most of the districts ingrowing types (i.e. with the exception of London, the lessurban district types) were in growing regions.

This might lead us to conclude that it is the growth ofthese less urban districts, at the expense of cities, that isdriving the regional population shift. However, if this werethe sole explanation, we would expect to see rural districts,mixed urban and rural districts, small cities and othergrowing types of districts gaining population or at leastholding their populations even in declining regions,gaining population from declining cities within their own

TABLE 8: Regional Population Change in the 1980s and 1990s

LondonSouth EastSouth WestEast of EnglandEast MidlandsWest MidlandsYorks/HumberNorth EastNorth WestWalesScotlandGreat Britain

Mid 1981population

(000s)68067245438148543853518749182636694028135180

54815

Mid 1991population

(000s)68297629468851214011523049362587684328735083

55831

Mid 2001population

(000s)73088022493754024183528349712519676729085064

57363

% Change1981-1991

0.4%5.3%7.0%5.5%4.1%0.8%0.4%

-1.9%-1.4%2.1%

-1.9%1.9%

% Change1991-2001

7.0%5.1%5.3%5.5%4.3%1.0%0.7%

-2.6%-1.1%1.2%

-0.4%2.7%

Share of total change1981-2001

2%38%30%26%16%4%2%

-5%-10%

6%-10%

Share of total change1991-2001

31%26%16%18%11%3%2%

-4%-5%2%

-1%


FIGURE 5: Regional Trends in the 1980s and 1990s

FIGURE 6: Distribution of District Type by Type of Region

region and in other regions.This was not entirely the case.Declining regions had a complete absence of high flyingdistricts, and few growers (Figure 7). Districts in theseregions were mainly ‘in the balance’ or declining, whatevertheir type. This suggests that although less urbaniseddistricts may be gaining population from declining citieswithin the region, they are not doing so at a fast enoughrate to offset their own population losses (arising from netout-migration or negative natural growth) and to sustainpopulation growth on any significant scale.

FIGURE 7: Population Change Rates, Region by Region(growing regions on the left of the graph, declining regions on the right)

Not all districts in declining regions declined. However,in general, types of district that were, across the countryas a whole, growing, were less likely to grow if they werein declining regions. Table 9 classifies population changeby region, only for the types of districts that, on average,grew during the 1990s (London, small cities, urban/rural,resort/port, rural and new towns). While there was ageneral pattern of more growth than decline in all regionsexcept the North East, it is evident that a much higherproportion of these districts grew in the growing regions(for example, East Midlands) than the declining ones(such as Scotland).The same pattern holds for industrialareas, which in the country as a whole grew more slowlythan the national average. Industrial areas in growingregions mainly grew, whereas industrial areas in decliningregions were more likely to decline (Table 10).

We do not have space here to show detailed analyses foreach region individually, but Maps 9 and 10, of the WestMidlands and North East, illustrate the general point. Bothshow the decline in the major conurbations that we reportedearlier. However, the West Midlands (a slightly growingregion) shows a general pattern of growth outside the


TABLE 9: Population Change for Districts in Growing District Types: Numbers of Districts in Each Category on Population Change Classification

LondonSouth EastSouth WestEast MidlandsEast of EnglandWalesWest MidlandsYorkshire and the HumberNorth EastNorth WestScotlandTOTAL

Decliners

1

3217

Slight decliners

11

21

1214

13

Minimal change

16

12121

23

19

Slight growers

61342653

431

44

Growers

1835319

253

107

87

153

High flyers

106779131

44

Total

3361421945111810

91616

280

Number of Districts

In the balance

TABLE 10: Population Change for Districts with Industrial Areas: Numbers of Districts in Each Category on Population Change Classification

LondonSouth EastSouth WestEast MidlandsEast of EnglandWalesWest MidlandsYorkshire and the HumberNorth EastNorth WestScotlandTOTAL

Decliners

3

1116

Slight decliners

41311431

18

Minimal change

1326

Slight growers

1

4

14

212

15

Growers

218123

14

22

High flyers

1

21

4

Total

41

1839819

126

71

Number of Districts

In the balance


conurbation, even in industrial districts. The North East (a declining region) by contrast, shows a general pattern ofdecline, with even rural districts showing only small growthor even decline.

These results suggest that what determines populationchange is not just what type of a place a district is (usingthis broad classification), but also where it is, in terms ofthe region in which it is located. Both intra-regional(between districts) and inter-regional factors seem to beat work. To fully disentangle these and to understandtheir relative weight, we need a more sophisticatedcategorisation of area types than is used here,defined by functional rather than administrativeboundaries, and classifying areas according to arange of variables including their region, location,size, industrial structure and population mix.8 Ouranalysis here merely serves to make the point thatdiscussions of the decline and renewal of cities cannotbe divorced from an understanding of their widerregional contexts.

MAP 9

MAP 10

8 An approach such as this was used by Champion et al. (1987),defining ‘functional regions’ and local labour market areas. The newONS classification might provide a similar basis.

Population change 1991-2001Decliners (more than 3% decline)Slight Decliners (-0.5% to -3%)Minimal Change (-0.5% to 0.5%)Slight Growers (0.5% to 3%)Growers (3% to 10%)High Flyers (10%+)

CONCLUSIONSThis census brief sets out the types of district making upthe urban areas of Great Britain. London is unique asthe capital with 13% of the total population (7.3million). Seven major ex-industrial conurbations, all inthe Midlands, North and Scotland, have 22% of thepopulation (12.5 million).The 14 other major cities (3.6million people) and 17 smaller cities (2.2 millionpeople) make up another 10%. Thus, 45% of the totalpopulation live in major urban areas. The 41 otherindustrial districts have nearly 8 million inhabitants(14% of the population) and the 24 new or expandedtowns, built since World War Two, have another 3 million(5%).The remaining third of the population (around 20million) lives in a mixture of smaller urban and ruralareas. Our work focuses on major urban areas and citiesin particular, although the interaction between differenttypes of built up area is strong - decline in one typemeans growth in another.

Overall the population of Great Britain grew by 2.7% inthe decade 1991-2001. 56% of all districts grew by morethan this; the strongest “growers” were the remoter, morerural areas, the mixed urban and rural areas and London.77% of all the fastest growing districts (more than 10%growth) were in these types of districts. Smaller citieswere also growing with only 7 exceptions.

Larger cities and metropolitan conurbations werealmost all declining or only expanding slightly. Amongthe seven major conurbations, the population of WestYorkshire expanded slightly - by 1% - while Merseysidesuffered the steepest decline - 5%. Overall, excludingLondon, the conurbations lost around 2% of their 1991population by 2001.

There is a strong overall pattern underlying the growthand decline. The industrial areas, concentrated in theNorth and West Midlands, as a whole, lost population;while the less industrial areas of the South West, SouthEast and Eastern regions gained. London and itssurrounding South Eastern region absorbed more thanhalf the total population growth of the country in the1990s - over 700,000 extra people. London stands outas the only major city to grow fast in the last decade,following several decades of steep decline, whichlevelled off in the 1980s.

Urban decline is about more than movement outwardsto the edge of cities. The 2001 census confirmed astrong tilting in population growth towards the SouthEast with a matching regional tilt in populationconcentration away from the North and Scotland.Behind these clear patterns lies a persistent regionaltrend that underlines the interaction between cities andtheir regions. Declining cities and districts are heavilyconcentrated in declining regions; growing cities areconcentrated in growing regions.

Population change reflects the significant continuingdecline of the most industrialised regions in the wake ofmanufacturing decline and steady growth in the lessindustrialised regions as new types of industry andservice-based enterprise grow more readily there. The

decline, though relatively small in percentage terms,becomes much more significant in the context of overallpopulation growth. Thus although there is measurable“de-urbanisation” going on, the decline of majorindustrial cities appears intimately linked with widerregional decline.

What explains these patterns? The Urban Task Force(DETR 1999), the Urban White Paper (DETR 2000)and the study of Scottish cities (2003), among manyother urban studies, set out a familiar history. Industrialcollapse left a legacy of high worklessness, poverty anddeclining social conditions. The physical environment ofindustrial areas was blighted by contaminated land,obsolete infrastructure and the debris of two centuries ofrapid growth and exploitation of natural local resources.Too little has been invested in environmentalreinstatement relative to these extensive damages andpoor social conditions caused by significant economicand environmental as well as social change, prove highlyresistant to improvement as the economy continues to“tilt away” from the declining industrial regions.

Meanwhile, the more diverse southern cities, generallylacking the extensive damage and sharp economiccollapse of the North and Midlands, are more attractivefor new investment and growth. As the investment scalescontinues to tilt in favour of these more attractive places,so the divide between North and South grows(Parkinson, 2003). New towns and growth areas in theSouth East are prospering, while the older industrialareas of the North face significant barriers to regrowingthe economy - most importantly their now largelyredundant industrial legacy.

Our analysis underlines this worrying trend.Yet it seemshard to imagine that the current pace of growth inLondon and South East will continue without someevening up of growth patterns. One obvious mechanismto increase the chances of this happening would be toinvest more in remediating the damaged urban landscapeof former industrial areas, particularly the major cities,which are the hub of their regional economies and thecentres of new economic activity. We will move on to thesigns of economic regrowth in the subsequent studies inthis series.

The papers that follow, looking at the ethnic compositionof populations, the distribution of economic activity,household, housing, tenure, and neighbourhood change,may shed more light on the powerful population trendsthat this first paper has highlighted. There is a generalacceptance that London and the South are the mostsuccessful parts of the economy; that the North andMidlands contain more deprivation, more sluggisheconomies, and are less generally successful. Our workunderlines the firm base on which these perceptions arefounded, while highlighting some variations in a strongpattern. The gaps in understanding are around the roleof successful regions in driving the national economy, themechanisms for preventing over-concentrations ofpopulation within an already highly congested region,and the potential for economic, social and environmentalbenefits of “re-tilting” growth towards currentlydeclining regions.



REFERENCESAtkins, D. et al. (1996), Urban trends in England: Latest evidence fromthe 1991 Census. London: DoE.Champion, A.G. et al. (1987), Changing Places: Britain’s Demographic,Economic and Social Complexion. London: Edward Arnold.DETR (1999), Towards an Urban Renaissance: Final Report of theUrban Task Force. London: DETR.DETR (2000), Our Towns and Cities: The Future: Delivering an UrbanRenaissance. London: DETR.Glaeser, E. and Shapiro, J. (2003), “City Growth: Which Places Grewand Why”, in Katz, B. and Lang, R. (eds.), Redefining Urban andSuburban America: Evidence from Census 2000. Volume 1. WashingtonDC: The Brookings Institution.Lupton, R. (2003), Poverty Street: The Dynamics of NeighbourhoodDecline and Renewal. Bristol: The Policy Press.Maclennan, D et al. (2003), Better Cities, New Challenges: A review ofScotland’s cities. Scottish Executive.Mumford, K. and Power, A (2003), East Enders: Family andCommunity in East London. Bristol: The Policy Press.Parkinson, M et al. (2003), EU non-capital cities project: Emergingfindings, trends, implications. Progress report to Core Cities WorkingGroup.Rogers, R. and Power. A (2000) Cities for a Small Country. London:Faber and Faber.Tunstall, R. (2004), Studying Urban Areas in the United States andUnited Kingdom using the Censuses 1980/01-2000/01.Washington DC:Brookings Institution, Center on Urban and Metropolitan Policy.

Publications arising from work at the Brookings Institution on the USCensus can be found at:http://www.brookings.edu/es/urban/issues/demographics/demographics.htm

sticerd.lse.ac.uk/dps/case/cbcb/census1.pdfsticerd.lse.ac.uk/dps/case/cbcb/census1.pdf ·...

Documents